论文部分内容阅读
针对轧钢生产过程中在线故障检测和故障诊断的问题,提出一种基于多核学习算法的钢铁生产轧钢过程在线故障检测模型.首先针对学习样本建立核主成分分析与支持向量数据域描述模型,然后基于T2、Q统计量,以及数据域描述包络情况对轧钢过程进行初步识别,最后构建基于多分类多核最小二乘支持向量机预测模型,对初识结果进行细分类,识别故障级别.利用上述模型对轧钢加热炉故障和机组故障进行了试验.结果表明,该方法能有效检测钢铁生产轧钢过程的故障.
Aiming at the problem of on-line fault detection and fault diagnosis in the process of rolling steel production, an on-line fault detection model based on multi-core learning algorithm for steel production during rolling process is proposed.Firstly, a principal component analysis and support vector data domain description model are established for learning samples, T2, Q statistic, and the data field to describe the envelopment of the rolling process, and finally build a predictive model based on multi-classification and multi-core least squares support vector machine to classify the initial results and identify the fault level.Using the above model The failure of rolling steel heating furnace and unit failure were tested.The results show that this method can effectively detect the failure of steel production and rolling process.